Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1777155

Summary: [DOCS] OpenShift 4.x Updating Cluster documentation should suggest backup of control plane
Product: OpenShift Container Platform Reporter: Joel Pearson <japearson>
Component: DocumentationAssignee: Kathryn Alexander <kalexand>
Status: CLOSED CURRENTRELEASE QA Contact: Xiaoli Tian <xtian>
Severity: low Docs Contact: Vikram Goyal <vigoyal>
Priority: low    
Version: 4.2.0CC: aos-bugs, jokerman
Target Milestone: ---   
Target Release: 4.2.z   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2019-11-28 00:21:42 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Joel Pearson 2019-11-27 03:17:31 UTC
Document URL: 
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-between-minor.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-cli.html

Section Number and Name: Prerequisites

Describe the issue: 
When upgrading a cluster in OpenShift 4.x the documentation doesn't suggest that you should perform a backup of the control plane first. On the first page: https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-between-minor.html it does mention that you can't perform a revert or rollback to a previous version.

Suggestions for improvement: 
In the prerequisites section it would be a good idea to suggest that a backup of the control plane has been performed.  Because if an upgrade fails, there isn't a supported rollback strategy without restoring etcd.

Additional information:

Comment 1 Vikram Goyal 2019-11-27 03:24:56 UTC
Hey Joel.

Thanks for submitting this bug.

I am not sure about this suggestion since as you correctly point out, you can't revert or rollback. Do you have any other info regarding this situation?

Comment 2 Joel Pearson 2019-11-27 03:47:53 UTC
Hi Vikram,

Clayton Coleman on the mailing list mentioned that I should raise a documentation bug about this:
https://lists.openshift.redhat.com/openshift-archives/users/2019-November/msg00068.html

The context surrounding this, is that I had a failed update from 4.2.2 to 4.2.4 and the cluster was no longer usable, and I didn't have a backup of the control plane. And the documentation says it is not supported to downgrade back to 4.2.2 (a downgrade might have actually worked, but I didn't do it properly).

My understanding is that to downgrade the cluster you need to restore the cluster to a previous state by using this: https://docs.openshift.com/container-platform/4.2/backup_and_restore/disaster_recovery/scenario-2-restoring-cluster-state.html

In terms of a suggestion:

For example on this page:  https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-between-minor.html#update-upgrading-web_updating-cluster-between-minor

This text:

----

Updating a cluster by using the web console
If updates are available, you can update your cluster from the web console.

You can find information about available OpenShift Container Platform advisories and updates in the errata section of the Customer Portal.

Prerequisites
Have access to the web console as a user with admin privileges.
----

Could possibly be changed to something like:

----

Updating a cluster by using the web console
If updates are available, you can update your cluster from the web console.

You can find information about available OpenShift Container Platform advisories and updates in the errata section of the Customer Portal.

Prerequisites
Have access to the web console as a user with admin privileges.
A backup (https://docs.openshift.com/container-platform/4.2/backup_and_restore/backing-up-etcd.html) has been performed in the event a rollback (https://docs.openshift.com/container-platform/4.2/backup_and_restore/disaster_recovery/scenario-2-restoring-cluster-state.html) is required
----

Then add that same backup reference to all 4 pages in the Prerequisite section:

https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-between-minor.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-cli.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-rhel-compute.html

What do you think?

Comment 3 Vikram Goyal 2019-11-27 03:51:08 UTC
Thanks for the added context Joel.

Let me work with the team to figure out the proper update for this. Since it is Thanksgiving week, an update might not be possible till next week.

Comment 4 Kathryn Alexander 2019-11-27 20:20:39 UTC
Here's a PR to apply the prereq to each of the update assemblies: https://github.com/openshift/openshift-docs/pull/18351

Comment 5 Vikram Goyal 2019-11-28 00:21:42 UTC
Updates were merged and pushed live. For example:

https://docs.openshift.com/container-platform/4.2/updating/updating-cluster.html
https://docs.openshift.com/container-platform/4.2/updating/updating-cluster-cli.html

Closing this bug as CURRENTRELEASE.